Machine Learning (ML) algorithms are powerful data-driven tools for approximating highdimensional or non-linear nuisance functions which are useful in practice because the true functional form of the predictors is ex-ante unknown. In this paper, we develop estimators of policy interventions from panel data which allow for non-linear effects of the confounding regressors, and investigate the performance of these estimators using three well-known ML algorithms, specifically, LASSO, classification and regression trees, and random forests. We use Double Machine Learning (DML) (Chernozhukov et al., 2018) for the estimation of causal effects of homogeneous treatments with unobserved individual heterogeneity (fixed effects) and no unobserved confounding by extending Robinson (1988)'s partially linear regression model. We develop three alternative approaches for handling unobserved individual heterogeneity based on extending the within-group estimator, first-difference estimator, and correlated random effect estimator (Mundlak, 1978) for non-linear models. Using Monte Carlo simulations, we find that conventional least squares estimators can perform well even if the data generating process is nonlinear, but there are substantial performance gains in terms of bias reduction under a process where the true effect of the regressors is non-linear and discontinuous. However, for the same scenarios, we also find - despite extensive hyperparameter tuning - inference to be problematic for both tree-based learners because these lead to highly non-normal estimator distributions and the estimator variance being severely under-estimated. This contradicts the performance of trees in other circumstances and requires further investigation. Finally, we provide an illustrative example of DML for observational panel data showing the impact of the introduction of the national minimum wage in the UK.

Fox News Flash top headlines are here. Check out what's clicking on Foxnews.com. For the first time, artificial intelligence has been used to deliver a speech by Pakistan's imprisoned former Prime Minister Imran Khan to supporters. The stunning development could help his political party to win upcoming elections, analysts said Monday. The replicated voice of Pakistan's most popular opposition figure was used to address a virtual event on social media watched by more than a million people.

Islamabad, Pakistan – Jailed former Pakistani Prime Minister Imran Khan has delivered a speech using an audio clip generated through artificial intelligence (AI) to address a virtual rally – the first event of its kind in the South Asian country. Khan delivered a four-minute address on Sunday night, using the clip, which was laid over a video containing his AI-generated image as well as photos from previous Pakistan Tehreek-e-Insaaf (PTI) rallies and earlier speeches by him. The PTI said its virtual rally drew more than five million views on social media platforms including YouTube, Facebook and Twitter despite internet outages reported from various parts of the country. The PTI organised the internet rally to bypass a government ban on public rallies by the party as it gears up for general elections, scheduled on February 8. "Our party is not allowed to hold public rallies. Our people are being kidnapped and their families are being harassed," the AI-generated voice mimicking Khan said in the clip, carrying a disclaimer that the speech is based on his notes from prison.

Adobe has abandoned its $20bn (£15.8bn) The Photoshop owner, which dominates the market with products including Illustrator and Acrobat Reader, said the two companies had come to a joint assessment that there was "no clear path" to regulatory approval. The UK's Competition and Markets Authority said last month that the deal would threaten competition in the product design, image editing and illustration markets. "There is no clear path to receive necessary regulatory approvals from the European Commission and the UK Competition and Markets Authority," the companies said in the joint statement. "Adobe and Figma strongly disagree with the recent regulatory findings, but we believe it is in our respective best interests to move forward independently," the chair and chief executive of Adobe, Shantanu Narayen, said.

Over the summer, a political action committee (PAC) supporting Florida governor and presidential hopeful Ron DeSantis uploaded a video of former president Donald Trump on YouTube in which he appeared to attack Iowa governor Kim Reynolds. It wasn't exactly real--though the text was taken from one of Trump's tweets, the voice used in the ad was AI-generated. The video was subsequently removed, but it has spurred questions about the role generative AI will play in the 2024 elections in the US and around the world. While platforms and politicians are focusing on deepfakes--AI-generated content that might depict a real person saying something they didn't or an entirely fake person--experts told WIRED there's a lot more at stake. Long before generative AI became widely available, people were making "cheapfakes" or "shallowfakes."

The White House has privately addressed concerns over an increasingly close relationship between Beijing and private industry in the Middle East that could see Chinese influence over powerful new artificial intelligence (AI) models. "It's very reminiscent of the Huawei issue where you have these technologies with 5G," Dr. Georgianna Shea, the chief technologist at the Foundation for Defense of Democracy's Center on Cyber and Technology Innovation, told Fox News Digital. "Everyone's using [5G], so that it becomes a backdoor into a lot of different systems within the United States," Shea said. "AI offers that same opportunity when [China] partners with our allies: They can both get in on the development side of it and, possibly, skew some of those biases or directly go through and pull out the intellectual property from what's being put into the model." The Biden administration has made clear in private discussions with the United Arab Emirates (UAE) that the oil-rich nation should pay close attention to ties between Beijing and the Emirati company G42, which launched its Jais AI model – reportedly the most advanced Arab-language AI model.

EXCLUSIVE: A top House Republican is warning that the U.S. needs to stay ahead of China, Russia and other adversaries in the race to dominate the artificial intelligence (AI) space, particularly with regard to the military. "We've got to develop it. It's got to be managed," Rep. Gary Palmer, R-Ala., chairman of the House Republican Policy Committee, told Fox News Digital when asked how the U.S. military could lead the AI sphere. Palmer suggested the integration of AI with quantum computing would be a significant part of military development going forward. "What that does just by itself – the ability to analyze a situation on the ground or in the air and have an almost instantaneous countermeasure or attack. That's what quantum computing does," Palmer said.

Fox News Flash top headlines are here. Check out what's clicking on Foxnews.com. Billionaire Elon Musk and OpenAI CEO Sam Altman are engaged in a raging battle over how much access the public should have to the technology behind artificial intelligence or AI. The most influential player in this space, however, is not Musk or Altman. A careful dusting for fingerprints reveals that it is none other than the president of the United States himself, Joe Biden.

The fundamental goal of small molecule discovery is to generate chemicals with target functionality. While this often proceeds through structure-based methods, we set out to investigate the practicality of orthogonal methods that leverage the extensive corpus of chemical literature. We hypothesize that a sufficiently large text-derived chemical function dataset would mirror the actual landscape of chemical functionality. Such a landscape would implicitly capture complex physical and biological interactions given that chemical function arises from both a molecule's structure and its interacting partners. To evaluate this hypothesis, we built a Chemical Function (CheF) dataset of patent-derived functional labels. This dataset, comprising 631K molecule-function pairs, was created using an LLM- and embedding-based method to obtain functional labels for approximately 100K molecules from their corresponding 188K unique patents. We carry out a series of analyses demonstrating that the CheF dataset contains a semantically coherent textual representation of the functional landscape congruent with chemical structural relationships, thus approximating the actual chemical function landscape. We then demonstrate that this text-based functional landscape can be leveraged to identify drugs with target functionality using a model able to predict functional profiles from structure alone. We believe that functional label-guided molecular discovery may serve as an orthogonal approach to traditional structure-based methods in the pursuit of designing novel functional molecules.

As AI systems proliferate in society, the AI community is increasingly preoccupied with the concept of AI Safety, namely the prevention of failures due to accidents that arise from an unanticipated departure of a system's behavior from designer intent in AI deployment. We demonstrate through an analysis of real world cases of such incidents that although current vocabulary captures a range of the encountered issues of AI deployment, an expanded socio-technical framing will be required for a more complete understanding of how AI systems and implemented safety mechanisms fail and succeed in real life. The rapid adoption and widespread experimentation and deployment of AI systems has triggered a variety of failures. Some are catastrophic and visible such as in the case of fatal crashes involving autonomous vehicles. Other failures are much more subtle and pernicious, such as the development of new forms of addiction to personalized content.